IJSTE - International Journal of Science Technology & Engineering | Volume 1 | Issue 3 | September 2014

ISSN(online) : 2349-784X

Automated Detection of Vascular Abnormalities

in Diabetic Retinopathy using Morphological
Entropic Thresholding with Preprocessing
Median Fitter
G. Supraja
M.Tech Student
Department of Electronics and Communication Engineering
Bapatla Engineering College, Bapatla, Andhra Pradesh, India

M. Baby
Assistant Professor
Department of Electronics and Communication Engineering
Bapatla Engineering College, Bapatla, Andhra Pradesh, India

Abstract
Diabetic retinopathy is one of the serious eye diseases that can cause blindness and vision loss. The complication of the diabetes
associated to retina of the eye is DR. A Patient with the disease has to undergo periodic screening of eye. For the diagnosis,
ophthalmologists use color retinal images of a patient acquired from digital fundus camera. Ocular funds image can provide
information on pathological changes caused by local ocular diseases and early signs of certain systemic diseases. The present
study is aimed at developing an automatic system for the extraction of normal and abnormal features in color retinal images. The
Preprocessing median fitter is applied before the Morphology. Morphological filter is tuned to match that part of vessel to be
extracted in a green channel image. To classify the pixels into vessels and non-vessels local thresholding based on gray level cooccurrence matrix is applied. The performance of the method is evaluated on two publicly available retinal databases with hand
labeled ground truths. The performance of retinal vessels on DRIVE database, sensitivity 91% accompanied by specificity of
94%. While for STARE database proposed method sensitivity 92% and specificity 90%. The system could assist the
ophthalmologists, to detect the signs of diabetic retinopathy in the early stage, for a better treatment plan and to improve the
vision related quality of life.
Keywords: Vessel Segmentation, Image Processing, Diabetic Retinopathy, Preprocessing Median filter, Morphological
filtering
________________________________________________________________________________________________________

I. INTRODUCTION
Diabetic Retinopathy (DR) is an eye disease which occurs due to diabetes. It damages the small blood vessels in the retina
resulting in loss of vision. Automatic segmentation of blood vessels in retinal funds image plays an important role in the
diagnosis of several pathologies, like hypertension, diabetes, and cardiovascular disease. Several morphological features of veins
and arteries (e.g. diameter, length, branching angle, and tortuosity) have diagnosis relevance. Accurate vasculature segmentation
is a difficult task for several reasons: The presence of noise, the low contrast between vessels and background, and the variability
of vessels width, brightness, and shape. Moreover due to the presence of lesions, exudates, and other pathological effects, the
image may have large abnormal regions. The risk of the disease increases with age and therefore, middle aged and older
diabetics are prone to Diabetic Retinopathy.
Retinopathy is a progressive disease, which can advance from mild stage to progressive stage. There are three stages based
on The National Screening Committee (NSC): (I) early stage or non-proliferate diabetic retinopathy (NPDR) or background
retinopathy, (ii) maculopathy and (iii) progressive or proliferate retinopathy. The early stage is further classified as mild NPDR
and moderate to severe NPDR [1], [2].In mild NPDR, signs such as microaneursyms, dot and blot haemorrhages and hard or
intra retinal exudates are seen in the retinal images. Microaneury are small, round and dark , red dots with sharp margins and are
often temporal to macula [1], [3]. Their size ranges from 20 to 200 microns i.e., less than 1/12 th the diameter of an average optic
disc and are first detectable signs of retinopathy. Haemorrhages are of two types: Flame and Dot-blot haemorrhages. Flame
haemorrhages occur at the nerve fibers and they originate from precapillary arterioles, which located at the inner layer of the
retina [4]. Dot and blot haemorrhages are round, smaller than micro aneurysms and occur at the various levels of retina
especially at the venous end of capillaries. Hard exudates are shinny, irregularly shaped and found near prominent micro
aneurysms or at the edges of retinal edema. In early stage, the vision is rarely affected and the disease can be identified only by
regular dilated eye examinations [5].

All rights reserved by www.ijste.org

21

Automated Detection of Vascular Abnormalities in Diabetic Retinopathy using Morphological Entropic Thresholding with Preprocessing Median Fitter
(IJSTE/ Volume 1 / Issue 3 / 004)

Fig. 1: Main stages of Retinopathy with the disorder

Fundus images are used for diagnosis by trained clinicians to check for any abnormalities or any change in the retina. They are
captured by using special devices called ophthalmoscopes. A typical fundus image with its features marked is shown in the
Figure 1. Each pixel in the fundus image consist of three values namely red, green and blue, each value being quantized to 256
levels. Diabetic Maculopathy is a stage where fluid leaks out of damage vessels and accumulates at the center of the retina called
macula (which helps in seeing the details of the vision very clearly) causing permanent loss of vision. This water logging of the
macula area is called clinically significant macular edema which can be treated by laser treatment.
Proliferate diabetic retinopathy, which is defined as the growth of abnormal new vessels (neovascularization) on the inner
surface of the retina are divided into two categories: neovasculature of the optic disk

Fig. 2: Different stages of Diabetic retinopathy

Diabetic Maculopathy is a stage where fluid leaks out of damaged vessels.

II. DETECTION OF BLOOD VESSELS

There are different images-processing methods that can be used for capturing variation. Methods include image segmentation,
edge or boundary detection, shape and texture analysis. The detection process can be carried out either on the original image or
in the transform domain. Some of the transform, Fourier transform, And discrete cosine transforms (DCT). This paper utilizes
preprocessing morphological filter for automated detection and classification of retinal images.
A. Preprocessing
In this pre-processing is used for better visualization. Image pre-processing can significantly increase the reliability of an optical
inspection. Several filter operations which intensity or reduce certain image details enable an easier or faster evaluation. In this
work pre-processing is done with morphological filter for better visualization. Median Filter in images finds the median pixel
value within the diameter that specified. It removes bright or dim features. Median filters are very effective in removing salt and
pepper and impulse noise while retaining image details because they do not depend on values which are significantly different
from typical values in the neighborhood. Median filters work in successive image windows in a fashion similar to linear filter. It
sorts all the pixels in an increasing order and takes the middle one. If the number of pixels is even, the median is taken as the
average of the middle two pixels after sorting.
B. Morphological Filter
The morphological operations include dilation, erosion, opening, closing etc.

All rights reserved by www.ijste.org

22

Automated Detection of Vascular Abnormalities in Diabetic Retinopathy using Morphological Entropic Thresholding with Preprocessing Median Fitter
(IJSTE/ Volume 1 / Issue 3 / 004)

1)Dilation
Dilation is a process that thickens objects in a binary image. The extent of this thickening is controlled by the Structuring
Element (SE) which is represented by a matrix of 0s and 1s mathematically, dilation operation can be written in set notation as
below
{ |( )
}
Where is an empty element and as is the structuring element. The dilation of A by As is the set consisting of all structuring
element origin locations where the reflected and transmitted As overlaps at least some portions of A. Dilation operation is
commutative and associative.
2) Erosion
Erosion shrinks or thins the objects in a binary image by the use of structuring element. The mathematical representation of
erosion is as shown below.
{ |( )
}
Erosion is performed in MATLAB using the command imerode (Image Name, SE).
3) Opening and Closing:
In image processing, dilation and erosion are used most often and in various combinations. An image may be subjected to series
of dilations and or erosions using the same or different SE. The combination of this two principles leads to morphological image
opening and morphological image closing. Morphological opening can be described as an erosion operation followed by a
dilation operation. Morphological opening of image X by Y is denoted by X O Y, which is erosion of X by Y followed by
dilation of the result obtain by Y closing and opening
(
)
(
)
Morphological closing can also be described as dilation operation followed by erosion operation. Morphological Closing of
Image X by Y is denoted by X Y, which is dilation of X by Y followed by erosion of the result obtained by Y. Image opening
and image closing and are implemented in MATLAB by the use of imopen(image name) and imclose(image name) respectively

III. EXTRACTION OF VESSELS

The enhanced vessel segments in the morphological filter response image, an effective thresholding scheme is required. The
entropy based thresholding using gray levels co-occurrence matrix is employed. It computes optimal threshold by taking into
account the spatial distributions of gray levels that are embedded in the co-occurrence matrix. The GLCM contains information
on the distribution of gray level and edge information, as it is very useful in finding the threshold value. The gray level cooccurrence matrix is a
square matrix of the gray scale image I of spatial dimension
with gray levels in the range [0,
1L-1]. It is denoted by
matrix. The element of the matrix specifies the number of transitions between all pairs
) it
of gray level in a particular Way. For each image pixel at spatial co-ordinate (m, n) with its gray level specified by (
) (
) (
) and (
) The coconsiders its nearest four neighbouring pixels at locations of (
) (
occurrence matrix is formed by comparing gray level changes of (
) to its corresponding gray levels, (
) (
)and (
). Depending upon the ways in which the gray level follows gray level diafferent definitions
of co-occurrence matrix are possible. The co-occurrence matrix by considering horizontally right and vertically lower transitions
is given by

(
{
(

)
)

(
(

)
)

Where, the total number of transitions in the co-occurrence matrix, a desired transition probability from gray level to gray
level is obtained as follows

Fig. 3: Gray level co-occurrence matrix

All rights reserved by www.ijste.org

23

Automated Detection of Vascular Abnormalities in Diabetic Retinopathy using Morphological Entropic Thresholding with Preprocessing Median Fitter
(IJSTE/ Volume 1 / Issue 3 / 004)

IV. ENTROPY THRESHOLDING

Based on the gray level variations wit in or between objects and background, the gray quadrants. Let
be the threshold within
the range
that partitions the gray level co-occurrence matrix into quadrants, namely A, B, C and D.

Fig. 4: Four quadrants of co-occurrence matrix

Quadrant A represents gray level transition within the objects while quadrant C represents gray level transition with in the
background. The gray level transition between the object and the background or across the objects boundary is placed in
quadrant B and quadrant D. These four regions can be further grouped into two classes, referred to as local and joint quadrant.
Local quadrant is referred to quadrant A and C as the gray level transition arises within the objects or the back ground of the
image. Then quadrant Band D is referred as joint quadrant because the gray level transition occurs between the object and the
background of the image.
The local entropic threshold is calculated considering only quadrant A and C. The probabilities of object class and background
class are defined as the normalized probabilities of the object class and background class are functions of threshold vector
( , ) are defined as

The second order entropy of the object is given by

( )

The local transition entropy A denoted by

( ).Similarly, the second-order entropy of the background is given by

( )

Up the local transition entropies, the total second-order local entropy of the objects and the background is given by
( )
( )
( )
Finally,
the gray level corresponding to the maximum of ( ) over
gives the optimal threshold for value
[
]
It can be seen that there exist small unconnected pixels in the threshold image. These isolated pixels are removed by
performing length filtering based on connected pixel labelling. The result of removing these unconnected pixels can be seen in
the final segmented image. To ensure that only the section of the image containing data is considered during image processing
and analysis, a mask image is generated for each image. It is applied to remove any artefacts present outside the region of
interest.

Fig. 5: Segmented vessels; Threshold response image, final segmented image after removing unconnected pixels.

V. RESULT
The retinal images from the DRIVE database and STARE database segmentation method. The manually segmented vessels
provided in both the databases are used as gold standard Figure and Figure illustrates the result of vessel segmentation on one of

All rights reserved by www.ijste.org

24

Automated Detection of Vascular Abnormalities in Diabetic Retinopathy using Morphological Entropic Thresholding with Preprocessing Median Fitter
(IJSTE/ Volume 1 / Issue 3 / 004)

the images in each database. The entire process of segmenting vessels was performance on Intel PC with 1.66GHz CPU and
512MB memory using MATLAB7.10. The processing of each image including convolution and thresholding took about 30
seconds. Morphological filtering is used to enhance the multi-oriented vessels. True positive are pixels Marked as vessels in both
the segmentation given by a method and the manual segmentation used as ground truth. False positives are pixels marked as
vessels as by the method, but that are actually negatives are pixels marked as background in both images. And false negatives are
pixels marked as background by the method, but actually are vessels pixels
Tab1e 1
Performance analysis using Ground Truth Table
Ground truth
Positive
Negative
Method result Positive
True positive False
Positive
Negative True
False Negative
Negative

From these sensitivity and specificity evaluated. Sensitivity gives the percentage of pixels correctly classified as vessels by the
method and specificity gives the percentage of non-vessels pixels classified as non-vessels by the method as follows
Sensitivity =
Specificity =
Where
is true positive,
is true negative,
is false positive, and
is false negative at each pixel. The method is
compared with the matched filter based method of using the DRIVE database Table shows that preprocessing Morphological
filter is better in classification of vessels with less false positive fraction rate.
Table 2
Performance of retinal blood vessels segmentation method on DRIVE database
Method
Sensitivity (%) Specificity (%)
Proposed Method
91
94
Morphological
86
91
filter
Gabor filter
85
90

The STARE database and the result are depicted in Table. Here also the proposed method performs better with lower
specificity even in the presence of lesions in the abnormal images
Table 3
Comparison of vessel segmentation results on STARE database.
Method
Sensitivity (%) Specificity (%)
Proposed Method
92
90
Hoover et. Al
75
92

Fig. 6: DRIVE database; (a) Input image; (b) Manual segmentation by expert; (c) Automatic Segmentation by the method

Fig. 7: STARE database; (a) Input image; (b) Manual segmentation by expert ;( c) Automatic Segmentation by the method.

All rights reserved by www.ijste.org

25

Automated Detection of Vascular Abnormalities in Diabetic Retinopathy using Morphological Entropic Thresholding with Preprocessing Median Fitter
(IJSTE/ Volume 1 / Issue 3 / 004)

VI. CONCLUSION
In this paper the sensitivity and specificity can be determined for diabetic retinopathy effected image and normal fundus images.
The method Success for all images due to the presence of preprocessing technique. It was found that the number of miss
classified pixels was less compared to other filter methods.

REFERENCES
[1]
[2]
[3]
[4]
[5]
[6]

R. Klein, Diabetic retinopathy, Public health, vol. 17, pp.137-158, MAY 1996.
Dougherty G., Johnson M. J., and Weirs M., -Measurement of retinal vascular tortuosity and its application to retinal pathologies, Journal of Medical &
Biological Engineering & Computing, vol. 48, no.1.pp. 87-95, 2010.
J.G.O Shea and D.A Infeld, Screening and monitoring diabetic retinopathy, Birmingham and Midland Eye Centre, 1999
K J.Frank and J.P Diabetic. Clinical review of diabetic eye disease Primary care perspective, Southern Medical Journal, Vol. 89, no. 5, pp. 463-470, May
1996.
N.J.Lingel, Care of the patient with diabetic retinopathy, Pacific on-line Optometry Education
http://www.moorfields.nhs.uk/

All rights reserved by www.ijste.org

26